Major depressive disorder (MDD) affects at least 10 % of the world population and globally ranks as the second leading cause of years lost to disability. Approximately a third of depressed patients fail to achieve remission; therefore, for a marked number of depressed individuals, currents treatments are not adequate. Deep brain stimulation (DBS) of the subcallosal cingulate white matter (SCCWM) has had success in treating, treatment resistant depression (TRD), but this therapy still requires scientific evolutio before it can be considered a clinical therapy. The overall research objective is to determine the theoretically optimal stimulation parameters, and in the long-term, evaluate prospectively the clinical outcomes in de novo patients. The first aim is to develop image-based computational models of the electric field in SCCWM DBS. We will use magnetic resonance (MR) images to define the geometry and electrical properties of the head, and computed tomography scans to determine the location of the DBS array. Six patients will be modeled: two responders, two non-responders that became responders, and two non-responders. This aim will generate one of the most anatomical and electrically detailed DBS electric field models ever created, which will then represent a gold standard for defining the level of detail necessary for accurate models of DBS. The second aim is to evaluate the neural response to SCCWM DBS. We hypothesize that target and non- target white matter tracts will be found amongst forceps minor, the uncinate fasciculus, the cingulum bundle, and short midline fibers projecting to subcortical structures. Axons will be modeled using cable theory, the trajectory of the axons will be defined by conducting probabilistic tractrography on a diffusion-weighted MR image, and multivariate statistical analyses will be used to correlate fiber activation (or lack of activation) with a response to stimulation. The outcome of this aim will be the identification of potential target white matter pathways that are necessary and/or sufficient for eliciting an antidepressant effect when stimulated. The third aim is to identify stimulation parameters that optimize the theoretical efficiency and selectivity of SCCWM DBS. By efficient, we mean using the least amount of electrical energy to activate target neural elements, and by selective, we mean the ability to activate target neural elements over non-target elements. The space of possible electrode configurations and stimulus waveforms is too large to be tackled by brute- force, so we use a numerical optimization algorithm to optimize stimulation parameters in the six model patients. This aim will establish a metric for improving the efficacy of SCCWM DBS, which we will test in future work. Successful completion of this research will advance our understanding of how activation of certain cortical and/or subcortical fiber pathways can produce an antidepressant effect in patients with TRD.